Re: Host-based may be the way to go, but network controls areneccessary

[Iljitsch van Beijnum <iljitsch@muada.com>]

On Sat, 23 Nov 2002, Scott W Brim wrote:

> > Before CIDR there was no multihoming issue.  :-)

> This statement doesn't make sense to me.  There was multihoming before
> CIDR.  We had hierarchy arguments.  We had already overwhelmed some
> routers with RIB size (iirc the first was when we hit 108 routes) and
> route lookup time.  All the problems were there in potential.

I really want to put this thread to bed, as it doesn't help us much with
what we're doing here, but...

Pre-CIDR everyone who was connected announced their class A, B or C
blocks into the global routing table, whether they were multihomed or
single-homed. So there were issues (and they were similar in nature to
the ones we're trying to solve here), but they were unrelated to
multihoming.